Hook. A line flashed through the morning Abyss Digest that someone with an engineering background can't ignore: "August 7. Saudi Arabia, Turkey, and Pakistan signed a 'Joint Defence Agreement' in Mecca: armed attack on one party = attack on all (wording modeled on NATO Charter Article 5). There will be a permanent secretariat in Riyadh and a ministerial committee." I dug in to verify — and an hour later realized this was the third attempt in seventy-one years to assemble a Muslim analog of NATO. The first (Baghdad Pact 1955 → CENTO → dissolved 1979) failed for external reasons: first the Iraqi Revolution of 1958, then the Iranian Revolution of 1979. The second (IMCTC 2015, 43 countries) turned out to be essentially a "club without collective defense" — a structure for intelligence sharing, not an alliance with article 5. And now — the third. Only this time there's no common enemy. There's Iran, there's Israel, there are Houthis. But the main driver isn't them. The main driver is disappointment in the United States, which in February 2026 struck Iran without thinking about consequences for those living nearby. And in this lies the most interesting plot: the first Muslim defensive alliance born not because someone came, but because someone left.
On Friday, August 7, 2026, in Mecca — sacred to Islam — in the same building where Islamic summits of the Organization of Islamic Cooperation are held — three gathered. Recep Tayyip Erdoğan (president of Turkey, NATO's second-largest army after the US), Mohammed bin Salman (crown prince of Saudi Arabia, de facto ruler, world's largest oil exporter) and Shehbaz Sharif (prime minister of Pakistan, the only Muslim nuclear power). They put signatures on a document that CNN, BBC and Al Jazeera that same evening called the Mecca Joint Defence Agreement, Reuters called a joint defence deal, and Turkish officials clarified they call it the Mecca Joint Deterrence Agreement. (Al Jazeera, August 7, 2026).
The wording in the official communiqué — almost a verbatim retelling of Article 5 of the 1949 NATO Washington Treaty (the very article under which after September 11, 2001, the alliance for the first time in history brought its forces to combat readiness for one country):
"The agreement aims to strengthen collective deterrence against any act of aggression and establishes that any armed attack on any of the three states should be considered an attack on all of them. It also provides for expanded cooperation in all aspects of defense between the three states."
(Joint statement, cited by BBC, August 7, 2026; CNN, August 7, 2026).
What's important — and what almost nobody wrote about in the first hours: the text contains not a word about what specific obligations parties assume in case of attack. No automaticity of "bringing to combat readiness." No clause about "joint armed actions." Not even response timelines. A Turkish official told Reuters the pact was "purely defensive in nature." Pakistan's foreign ministry added diplomatic pathos about "historic ties of brotherhood and Islamic solidarity" (Al Jazeera, August 7, 2026).
Harlan Ullman, senior analyst at Killowen Group, said the key phrase to Al Jazeera that same day: "We have not seen the text of the Mecca Declaration. We assume it says 'an attack on one should be considered an attack on all' — and 'should be considered' is the operative word, because that does not mean 'must necessarily'" (Al Jazeera, August 7, 2026).
And that's already an admission in itself. An alliance that can work and an alliance that will work are two different engineering projects. The first is signed. The second — isn't.
The treaty was prepared for almost three years. Al Jazeera, citing Turkish officials, says directly: negotiations began shortly after October 7, 2023, after Hamas's attack on Israel and the start of the Israeli operation in Gaza. But they accelerated in six specific weeks of 2025–2026.
September 17, 2025. Three days before — September 9, 2025 — Israel carried out an airstrike on Doha, capital of Qatar, where a Hamas leadership meeting was taking place. This was the first-ever direct Israeli strike on the capital of a Gulf Cooperation Council member state. Eight days later Saudi Arabia and Pakistan signed their own bilateral mutual defence pact (Reuters, AP, Al Jazeera, September 17, 2025). For the first time a nuclear Muslim power officially stood to defend an oil monarchy. Nobody said this aloud, but in Riyadh they read it correctly: "Israel can strike any of us. And we need someone to shield us."
February 28, 2026. At 5:47 a.m. local time (UTC+3:30) Israel and the US began the largest air operation since 2003 — strikes on Tehran and other Iranian cities. Iran within hours delivered a counter-strike — not at Israel and not directly at American bases, but at six Arab states on whose territory these bases are located: UAE (Dubai, Palm Jumeirah, one dead, four injured), Bahrain (5th Fleet headquarters, hit on residential tower near headquarters), Kuwait (Ali Al Salem Air Base, drone at airport), Qatar (long-range detection radar in the north, intercepted "according to pre-approved security plan"), Jordan (indirectly), Saudi Arabia (Riyadh and eastern province, all attacks repelled, but the very fact). Iran's Islamic Revolutionary Guard Corps (IRGC) declared directly: "All American assets in the region are legitimate targets" (Al Jazeera, February 28, 2026).
April 8, 2026. First ceasefire between US and Iran, achieved through Pakistan's mediation (Chief of Army Staff Asim Munir personally flew to Tehran). But two months before that, March 19, foreign ministers of Turkey, Pakistan and Saudi Arabia already met in Riyadh on the sidelines of the Islamic Summit and for the first time publicly discussed "what a joint defense document should contain" (Al Jazeera, August 7, 2026).
August 7, 2026. Signing.
Six months. Three triggers: Doha strike (anyone can become a target), February War (the umbrella doesn't protect — it attracts strikes), Iranian response to others' bases (they hit us for the American umbrella). And into this corridor fit the signature under an alliance prepared three years but couldn't be assembled — until the US itself showed their guarantees don't work.
Christopher Coates Ulrichsen, research fellow at Baker Institute (Rice University), formulated it precisely: "States in the region have harboured growing doubts about the reliability of US partnership for more than a decade. And the decision to embark on military operations against Iran, and its subsequent conduct, and the periodic attempts to end it — all of this has only heightened those doubts and accelerated efforts to diversify security and defense relationships" (Al Jazeera, August 8, 2026).
Sina Toossi (Center for International Policy): "This is one of the first major geopolitical signs that the regional order is being restructured under the influence of the US-Israeli war with Iran. The significance is less in the formal military alliance and more in what it says about the reassessment of security by regional states themselves" (Al Jazeera, August 8, 2026).
And Andreas Krieg (King's College London) — to Al Jazeera, August 8, 2026: "For Riyadh, the experience of the current war has cemented an uncomfortable conclusion. American military power remains indispensable, but American relationships do not guarantee immunity from attack and do not necessarily mean Washington will act in Saudi Arabia's interests. So Riyadh wants greater strategic depth and greater choice."
To understand how serious (or not serious) the current pact is, we need to walk through two previous attempts. Both — failed, and both — left a clear trace in regional memory.
February 24, 1955 Iraq and Turkey signed the "Pact of Mutual Cooperation" in Baghdad — the first Muslim military alliance in history. Within a year Britain (April 5, 1955), Pakistan (September 23, 1955) and Shah Mohammed Reza Pahlavi's Iran (November 3, 1955) joined. Headquarters — in Baghdad. Name — Baghdad Pact (Wikipedia: Central Treaty Organization, Encyclopaedia Iranica, US State Department, Office of the Historian).
Why it arose. Same logic as NATO: deterrence. Only instead of the USSR as potential aggressor — the Soviet Union as real threat to the southern flank. Strategy known as "Northern Tier" — northern belt of anti-communist states along USSR's southern border (Turkey — Iran — Pakistan). John Foster Dulles, Eisenhower's secretary of state, lobbied the pact, but the US didn't formally join: "because of the pro-Israel lobby and difficulties with Congressional approval." The US joined the alliance's military committee only in 1958, when the Baghdad Pact was already beyond saving (Wikipedia, ibid).
Why it failed. Three blows in a row, all from outside.
First. July 14, 1958. Iraqi Revolution: army officers led by Abd al-Karim Qasem overthrow King Faisal II's monarchy, kill Prime Minister Nuri al-Said (who, incidentally, was the main Iraqi supporter of the pact). New government withdraws Iraq from Baghdad Pact, opens diplomatic relations with USSR, takes non-alignment position. Headquarters moves from Baghdad to Ankara (October 1958). Alliance loses a quarter of members and renames itself CENTO — Central Treaty Organization (August 19, 1959).
Second. 1965 and 1971. Pakistan requests CENTO's help in wars with India. Allies refuse. Official wording: "CENTO was created to deter USSR, not India." Pakistan loses East Bengal (Bangladesh, 1971). Inside Pakistani establishment CENTO forever gets branded "the club that watches you get cut" (Wikipedia, ibid).
Third. 1974. Turkey invades Cyprus. Britain withdraws its air squadrons (9th and 35th, Vulcan bombers), formally "reserved for the alliance." US Congress freezes military aid to Turkey — despite two presidential vetoes. Alliance exists only on paper.
Finale. March 11–16, 1979. Iranian Revolution. Iran's provisional government exits CENTO. March 12 — Pakistan. March 16 — official dissolution. CENTO's secretary general, Turkish diplomat Kamuran Gürün, announces convening a council for formal liquidation (The Gazette, Montreal, March 17, 1979, cited by Wikipedia).
The main lesson of the Baghdad Pact — it's not about communism. It's about reputation. An alliance that doesn't come to an ally's aid in trouble ceases to be an alliance faster than the adversary can deliver the next strike. CENTO outlived its own death in 1974 but existed on paper five more years because nobody wanted to admit the alliance died from its own indifference.
December 15, 2015 Saudi Arabia's crown prince Mohammed bin Salman (then defense minister, 30 years old) announced creation of Islamic Military Counter Terrorism Coalition (IMCTC) — "Islamic Military Coalition to Combat Terrorism." Initially 34 countries, by May 2025 — 43 states, including Egypt, Pakistan, Turkey, Jordan, Bahrain, UAE, Senegal, Benin, Chad, Niger, Burkina Faso, Cameroon, etc. Headquarters — Riyadh. Commander since 2017 — former Pakistan Army Chief of Staff Raheel Sharif (Wikipedia: Islamic Military Counter Terrorism Coalition).
Why it arose. In 2014–2015 the Islamic State (ISIS) controlled territory the size of Great Britain. Saudi Arabia wanted to lead the Islamic world in fighting terror — and simultaneously put Iran in the position of "not participating because not invited." Iran, Iraq and Syria (Shia governments) did not join the alliance. Analysts immediately called it a "Sunni sectarian coalition" (Hakeem Azameli, Iraqi parliament, cited Euronews, AFP, December 2015).
Why it didn't become a real alliance. Because it has no article 5. No obligations. No collective defense. There's intelligence sharing, joint exercises, coordination — but no more. After Jamal Khashoggi's murder in the Istanbul consulate of Saudi Arabia (October 2, 2018) trust in Riyadh in the Muslim world fell, and IMCTC essentially stopped expanding. Its real combat experience — Yemen intervention from 2015 — showed the alliance doesn't work for combat operations: Saudi aviation conducts the operation, allies give words of support.
Main lesson of IMCTC — it's about the difference between presence and participation. 43 countries in signatures, one country in battle. Alliance without automaticity isn't alliance, it's political declaration in memorandum form.
And here's the third attempt. Three countries, not forty-three. Article 5 in wording — but without automaticity in mechanics. Headquarters — again Riyadh, as in IMCTC. And for the first time — without USA, without Britain, without Iran.
This is true novelty. Neither the Baghdad Pact (which had London and tacit Washington), nor IMCTC (which had London's shadow and Pentagon's formal support) existed without Western support. The Mecca pact — the first purely Muslim alliance with not a single white person at the table, not a single Pentagon advisor, not a single line about NATO coordination. This is consequence, not cause. The cause — the United States stopped being a reliable guarantor, and now even their closest partners have to write their own insurance.
Now — to the engineering part. What the pact has, what it doesn't, and where its weakest nodes are.
First node — NATO incompatibility. Turkey — NATO member since 1952. NATO Article 5 obliges Ankara to defend any alliance member. Mecca pact Article 5 obliges Ankara to defend Riyadh and Islamabad. If these two obligations ever collide, Turkey has no choice procedure. And more importantly, NATO formally doesn't react, but in Brussels this pact is perceived as undermining their exclusivity. NATO analysts already compare the situation with 1960s debates about de Gaulle's "force de frappe" — independent nuclear factor inside alliance irritates headquarters even if it doesn't formally violate charter.
Second node — Pakistani-Indian. BBC directly poses the question: "Should Saudi Arabia declare war on India if conflict with Pakistan resumes?" (BBC, August 7, 2026). In May 2025 India and Pakistan already exchanged strikes for several days. Now the border is burning: Pakistan since February 2026 drawn into border conflict with Afghanistan (Taliban, new government). If something explodes in this corridor, Ankara will face a choice — and its choice isn't predetermined.
Third node — internal Saudi. Mohammed bin Salman isn't king. He's crown prince. King Salman bin Abdulaziz Al Saud (90+ years old) formally remains head of state. Inside royal family MBS has long had influential opponents, primarily those who remember how Sadat in 1972 signed Soviet-Egyptian friendship treaty and in 1979 was assassinated for it. If MBS loses power in two years, the pact may remain without signature — literally. CENTO had same dynamics play fatal role: Shah of Iran left, and alliance ended five days later.
Fourth node — sectarian. Alliance is officially Sunni. Iran — Shia. Inside Saudi Arabia — significant Shia minority (10–15%, mainly in Eastern Province where oil is). Inside Pakistan — significant Shia minority (10–15%, mainly in Balochistan and Karachi). Shia minorities are Iran's potential fifth column inside two alliance countries. Paradox: the stronger the pact, the more Iran will pressure Shia minorities, and the more Iran pressures, the more this pushes Riyadh and Islamabad toward each other. Vicious circle.
Fifth node — Russian and Chinese. Moscow and Beijing in 2025–2026 building their own mesh network (separate topic, discussed in previous longread about "axis worse than axis"). Mecca pact is parallel project for them, not necessarily coinciding. China, for example, isn't interested in escalation on its western border (Xinjiang bordering Pakistan), and Russia sells weapons to both Iran and Saudi Arabia simultaneously. If Moscow and Beijing decide Mecca interferes with their agenda, they have tools — from economic pressure on Pakistan (sitting on Chinese loans through CPEC) to diplomatic signals to Riyadh.
Trump publicly approved the pact on Truth Social signing day: "Very pleased, finally signed. Middle East coming together." Pakistani PM Sharif thanked Trump in response for "courage" in preventing "nuclear catastrophe in South Asia." At rhetoric level — Washington welcomes. At substance level — this is admission the region no longer wants to depend on Washington. Trump called it victory. In reality it's Washington's exit from exclusive guarantor role, packaged as praise.
Brett McGurk, key architect of Biden's Middle East policy, wrote on social media: "This is the era of tabula rasa in the Middle East — liquid, undefined." This is perhaps the most honest description of the moment (Al Jazeera, August 8, 2026).
Yasmine Farouk (International Crisis Group): "These countries — all are US partners, and they recognize there's no structure that can replace the United States" (Al Jazeera, August 8, 2026). Meaning — supplement, not replacement. But you have to start somewhere.
No official reaction. Neither Netanyahu's office nor foreign ministry commented on the pact in first 48 hours. This is unprecedented silence for a country whose foreign ministry press service comments on every sneeze in the region.
Al Jazeera interprets this as "Jerusalem taking pause to assess consequences" (August 8, 2026). Haaretz (Zvi Bar'el, August 9, 2026) formulates harsher: "New bloc of three countries doesn't necessarily create new military threat, but redraws region under itself — Trump's limited Middle East policy, weaker Israel with less influence in Washington, new regional position on Iran."
In February 2026 former PM Naftali Bennett already called Turkey "the new Iran" and accused Ankara of trying to "turn Saudi Arabia against Israel and build hostile Sunni axis with Pakistan." This prophecy came true six months later, and now Bennett in his own words looks like a prophet. Which is good for his political career but not for Israel.
No official reaction. Which itself is a reaction. Tehran doesn't recognize the pact, doesn't respond at government statement level. Only voice — Ebrahim Rezaei, member of Iran's parliamentary national security committee, on X: "Saudis should understand that a paper agreement with Turkey and Pakistan will not bring them security, just as decades of one-sided dependence on the Americans failed to do so" (Reuters, BBC, August 7, 2026).
This is mirror rhetoric. Iran accuses Riyadh of exactly what Riyadh accuses Washington: dependence on someone else's umbrella. And in both cases the argument has truth. Iranian argument is stronger than it seems: Central Treaty 1955 was also "paper" — until Iraqi revolution 1958 happened and pact collapsed.
But there's counter-argument Iranians don't mention: unlike Baghdad Pact, Mecca pact has no single sponsor that can be collapsed by removing it from participants list. Network of three major regional states is less vulnerable than network of five with one external guarantor (London + hidden Washington). Iran can't "withdraw" Turkey from pact like Qasem in 1958 "withdrew" Iraq from Baghdad.
UAE and Bahrain — don't comment publicly. But both countries suffered Iranian strikes February 28, 2026 alongside Saudi Arabia, and both have frictions with Riyadh (UAE — over Yemen and oil production; Bahrain — traditionally closer to Saudi Arabia but with own frictions). If these two countries joined the pact, its weight would grow radically. But so far — no.
Egypt — didn't sign. Yasmine Farouk in July 20, 2026 report (International Crisis Group) writes that "closer defence ties between Turkiye, Pakistan, Saudi Arabia and also Egypt reflects a hard-won realisation among these four states that security in their neighbourhood can no longer be left at the mercy of rivalry between the US and Israel, on one side, and Iran, on the other" (Al Jazeera, August 7, 2026). But Egypt is also not Iran's enemy (Iranian instructors fought in Syria alongside Russians, and Egypt knows this), and Cairo isn't ready to sign under "deter Iran" explicitly. This is future, not now.
Qatar — tried to be bridge between US and Iran in 2024–2025, lost trust after Doha strike September 9, 2025. Now — in suspended state. Al Jazeera doesn't accidentally call signatories "not implying Iran" but implying Israel and Houthis simultaneously — this allows Qatar to stay aside without losing face.
Since 1945 Middle East lived in American umbrella paradigm — first through "Eisenhower Doctrine" (1957), then through bilateral treaties with Saudi Arabia (1951, renewed repeatedly), Iran (until 1979), Kuwait, Bahrain, UAE, Qatar, Oman, Jordan. US guaranteed oil flow security in exchange for bases and petrodollars. This paradigm lived 80 years. Now it's ending — not because Washington decided to leave, but because partners for the first time signed paper where it's absent.
This is not Muslim NATO analog (as headlines love to write) — this is first independent defense triangle outside Western security system. If it lasts five years, others will follow: UAE + Bahrain + Egypt, or Algeria + Tunisia + Morocco, or Central Asian states. And then NATO and CENTO in their old form will stop being the only language of collective defense in the world.
Saudi Arabia stops being country that pays for security (through buying American weapons, maintaining American bases, paying Pentagon services). It becomes investor in regional security — co-founder of alliance, finance provider, headquarters host. This is historic shift in Riyadh's position, which in last century never played such role (Baghdad Pact Saudi Arabia didn't participate, in IMCTC participated as "one of 43" without leading position). Now — center.
Pakistan for decades lived in Indian focus — Kashmir problem, nuclear deterrence of India, Afghan rear. Mecca pact — Islamabad's first serious regional role outside Indian triangle. Pakistan — mediator between US and Iran (April 8, 2026), nuclear umbrella over Saudi Arabia, potential arbiter between Sunni states. This is new Pakistan, and it's not certain old Pakistan (with its internal political turbulence) can handle this.
Erdoğan — only NATO leader who simultaneously signed Mecca pact and openly conflicts with Israel (rhetoric level, but still). This is unprecedented position for country with NATO's second-largest contingent. Inside NATO there's quiet irritation: Ankara, as always, plays on two chessboards (Assad and Russia 2015–2024, now Israel and Iran). Mecca pact formalizes this. For first time Turkish "multi-vectorness" — not habit but institution.
Honestly, what I can say about Mecca pact: this isn't alliance of strength, it's alliance of fear. Three countries, each separately vulnerable to American inconsistency (Turkey — through conflict with Greece and Kurdish question; Saudi Arabia — through Iranian strikes; Pakistan — through India and Afghanistan), gathered together not because they have common victory, but because they have common fear: being left one-on-one with world where Washington no longer wants to be policeman.
And in this sense Mecca pact — mirror of Baghdad Pact 1955. That too was born not from common victory but from common fear of USSR. But that had external master — London and hidden Washington. This has nobody. And in this lies its vulnerability and strength simultaneously. Vulnerability — because without sponsor alliances converge to zero, like CENTO. Strength — because without sponsor alliance can't be collapsed by one decision from one uncle in Washington.
This is main shift of 2026: for first time in eighty years major Muslim states signed defense document where Washington is guest, not host. What to do with this — next attack will show. If real incident happens and three countries execute article 5 for first time — this is new regional reality. If incident happens and article 5 stays on paper — this is new Baghdad Pact 2.0.
Historical experience isn't on pact's side. Of four Muslim defensive alliances last 70 years (Arab League 1945, Baghdad Pact 1955, IMCTC 2015, Mecca pact 2026) none passed real crisis test. Arab League didn't defend Palestinians in 1948, 1967, 1973. Baghdad Pact didn't defend Pakistan in 1971. IMCTC didn't enter Yemen as unified force. Structurally — Mecca pact has better chances because inside it — three nuclear or near-nuclear potentials, two second regional armies and largest oil wallet. But until first crisis it's still claim, not result.
Pact's sharpest engineering problem — absence of common enemy. Baghdad Pact had USSR. IMCTC had ISIS. Mecca pact has Iran, Israel, Houthis, US-as-unreliable-ally, and Pakistani-Indian border. Five potential enemies — that's zero common enemies. Alliances hold not on "against whom" but on "for what." And until Mecca didn't say "for what" — we have formula but no answer.
And last. This story very quietly rewrote one of most persistent illusions of Middle East politics — that "Muslim world" isn't capable of formalized collective defense because "too fragmented" (Sunni-Shia, Arabs-Turks-Persians, monarchies-republics, rich-poor). Seventy-one years this illusion lived on strong foundation: Arab League failed everywhere it was tried. And here August 7, 2026 three countries, each of which previously would look at alliance idea skeptically, signed document. Not because they stopped being different. But because they understood: world around became so unstable that even bad alliance is better than none. This isn't optimism. This is insurance policy issued in world where insurer (US) stopped paying claims.
And last thing to add — about Israel's silence. When country whose foreign ministry press service responds to everything stays silent 48 hours after such document signing — this isn't pause. This is diagnosis. Israel for first time in thirty years found itself not at wheel of regional architecture but before its door. Door opened, they weren't invited. And Jerusalem's silence — most honest answer to question how deep this shift is.
Sources the longread relied on: