Takamul (تكامل) The region’s success comes from the close ties between government, innovators, and enterprises. Cross-sector alignment is no longer just a nice-to-have; it's the core engine driving regional growth. Here is how that integration played out across the MENA markets this week: 1. The Mohammed Bin Rashid Innovation Fund (@MBRIFund), the UAE Ministry of Finances (@MOFUAE) innovation initiative, partnered with the Dubai Future District Fund (@dubaifuturefund). MBRIF-backed startups get introduced to DFDFs network of fund managers, and DFDFs portfolio companies get access to MBRIFs federal programmes. @nadersalim , MD of DFDF: "Founders should not have to navigate federal and Dubai-level support separately." 2. @geterad closed a $22M Series A led by Middle East Venture Partners (@MEVPcapital), with @SVC_SA, @500GlobalVC and @ANB_Capital among the new investors. A government-backed fund of funds, a banks investment arm and regional VCs on one cap table. erad gives SMEs in Saudi and the UAE Shariah-compliant working capital, approved in about 48 hours. Congrats @SalemAbuHammourand team! 3. Oman launched a National Fintech Strategy led by the @centralbankoman, together with the Financial Services Authority (@fsa_oman), the Ministry of Finance and Invest Oman. The new Oman Fintech Gate gives local and foreign fintechs one entry point to learn the requirements and start their application. Four government bodies, one front door for founders. 4. @Investqa signed an MoU with Aurion Capital and LG NOVA, LG Electronics innovation centre, to explore tech investment in Qatar across AI, digital health and fintech. The partners expect a new hub at Qatar Science & Technology Park (@QSTP) to connect innovators with investors across the GCC. An investment agency, a corporate innovation arm and a capital partner at one table. And we still invest globally. 5. @ShorooqPartners and @G42ai’s @PresightAI joined @mavenrobotics ( the world’s leading general-purpose AI robots) $100M Series A in Silicon Valley. As @bilalabaloch of Shorooq put it, "this is how you invest in the frontier from the region, for the world." 6. @Microsoft plans more than $10B for cloud and AI across Kuwait, Qatar, Saudi Arabia and the UAE by 2030, with @G42ai, @HUMAIN and QAI as national partners.

k/acc knowledge acceleration coined by @beffjezos, who else. fav word of the week.

Impressive. This set of talks, along with the new America dot gov website, has the unmistakable feel of an Apple launch event. They're rebooting the federal government as a tech startup. Several observations. (1) First, with this new AI interface to the entire federal government, if it works as promised, there won't be a single American who can't say their life wasn't helped at least in part by AI, or that the benefits accrued solely to zillionaires. This in itself cannily mitigates anti-datacenter sentiment. (2) Second, Joe Gebbia and Marco Rubio's talks on simplifying name changes and passport filings introduce a completely new era where the US government is actually launching tangible features that benefit users. (3) Third, this launch event in and of itself may visibly drive reform. It could become much more important than the State of the Union. Unlike laws, code is user-visible, and can be instantly audited by the public, as Obama found out with the failed obamacare website launch. So: the annual America dot gov event may be substance, whereas SOTU is traditionally optics. I wouldn't be surprised to see the President present at the next one. (4) Fourth, this kind of event is a perfect example of where tech and red cooperate best. They both hate bureaucracy. Automating paperwork with code reduces the power of statists over both capitalists and nationalists. It's also hard to undo by other admins: once you digitize a process, hard to turn it back to paper. (5) Fifth, it's an obvious point perhaps, but think about how much influence YCombinator has had in 20 years. From Gebbia's Airbnb joining their 2009 batch to a complete overhaul of the user interface of the US bureaucracy, and thus the US government. (6) Sixth, this is another take on the network state. The functional part of the US (the network, meaning tech) is merging with and partially reforming the dysfunctional part (the state, meaning bureaucracy) by automating 20th century paper forms. As a tech startup, there will no doubt be bugs and user issues and so on, but unlike with paper laws those issues will be visible and the code fixes will be verifiable. A new era of AI America. Rooting for them.

Okay I got to have to admit you’re all right. Elon fixed the feed.

Gm. Have a good start of a new week. Keep going.

Having fun is my competitive advantage 😈 PC: @ashebytes on an unnamed epic film camera

𝐓𝐡𝐞 𝐨𝐧𝐥𝐲 𝐦𝐨𝐚𝐭 𝐥𝐞𝐟𝐭 𝐢𝐬 𝐭𝐡𝐞 𝐜𝐨𝐧𝐯𝐞𝐫𝐠𝐞𝐧𝐜𝐞 𝐨𝐟 𝐛𝐢𝐭𝐬 𝐚𝐧𝐝 𝐚𝐭𝐨𝐦𝐬. Wait, what? In July, I shipped a video as part of my work at @ankhlabs about which moats actually survive in a "post-software" world. The thesis: hardware, network effects, regulatory depth, frontier innovation, brand, owning distribution or owning data is what's left. Today, I'd like to double click on the hardware moat, how the defensibility moves to physical reality: the atoms. Bits (Software) • Asset: Apps, models, workflows, dashboards • Time to copy: Days • Risk: A frontier lab ships the feature or an AI agent rebuilds it overnight Atoms (Physical World) • Asset: Robots, chips, dark stores, fleets, clinics, power plants • Time to copy: Years • Risk: Someone has to physically rebuild supply chains and infrastructure Look at where convergence of atoms & bits hold defensibility: • Device atoms: Hardware in homes or wearables (@xpanceo, @Figure_robot , @ouraring ) • Network atoms: Physical assets & logistics (@snoonu_qa fleet, dark store networks) • Permissioned atoms: Physical + regulatory depth (@geidea 700,000 terminals behind the first non-bank acquiring license SAMA has issued) What got me to this framing was last week's @ycombinator's read on startups in 2026, and their shift in acceptance numbers show exactly where venture is heading: • Robotics: 1% → 6-7% • Industrial Manufacturing: 4% → 10% • Defense Tech: 1.5% → 5% • Semiconductors & Photonics: 1% → 4% • Energy Infrastructure: 1% → 3% The founder profile shifted with the category. 1 in 6 holds a PhD, and founders in their 30s, 40s, and 50s with deep domain experience in heavy industry, biology, and logistics are becoming the norm. I'm curious to see how these researchers turn into serial entrepreneurs the next years. My take on how this shifts the startup playbook: • GTM changes: Success now requires regulatory depth, governmental buy-in, enterprise integration, and robust infrastructure from Day 1 rather than relying on cheap customer acquisition loops and treating regulation as an afterthought. • Brand as a force multiplier: Physical infrastructure provides the initial moat, but long-term trust, community, and brand are what prevent commoditization over time. So while the convergence of bits and atoms isn't the only surviving moat, scaling it requires far more than a simple community launch and brand remains your most powerful compounding moat in the long run.

Industrial Revolution: replaced human muscle Cognitive revolution: replacing intelligence The constant: relationships you form today that will stay.

IMO the most interesting thing that happened in society today: The WSJ's leading op-ed today, by Stan Druckenmiller, was entirely AI generated. If you read the text, it's glaringly obvious. There's no way the editors at the WSJ don't have an ear for AI-written prose. They know what AI sounds like; they're not stupid. They published this anyway. Is this the new normal? Does it mean AI writing is finally good enough to publish in the WSJ, and we need to stop being so precious about AI-voice? Or does it mean that if you're famous and important, it's OK to just push out AI-written content so long as your byline is on it and you stand by the consequences? We understand that famous people have speechwriters, and so long as they deliver the speech, we consider it theirs. Maybe that feeling that something AI written is not your authorship is an anachronism that will fade away in a few years.

Wow. Elon just released the ENTIRE X algorithm. Nothing held back I just went through all 300,000+ lines of code What I read blew me away Here’s everything you need to know about how to go viral and if you can still get shadowbanned: 🧵

“Everything seems fake now.” .@lulumeservey on companies paying for retweets on their product launches: “Every launch uses the same slop cinematic video, same template.” “Half of them, you go to the engagement—it's completely fake.” “You can just start to tell that the amount of engagement this thing has is completely disproportionate to how interesting it actually is. And when you look through the people engaging with it, it's all some random account in a developing country that has no other tweets other than shilling products like this.” “When that happens, I write off the entire company as fake.” “I don't think, ‘This was a fake video or fake engagement.’ I think to myself, ‘It's a fake company with a fake founder. I will never buy their product or do business with them. I would never recommend somebody go work with them. The whole thing is fake.’” “It's better to put out something that truly represents the best version of what you’re doing that might get lower engagement, but the people who are engaging are actually real people, and they care. It's better to get 50 people who feel like this is speaking to them than to get 4,700 retweets that seem obviously fake.”

Had fun the last weeks working on the digital asset compliance landscape with the wonderful @makhavannik and @0xAlaz_ . Having been working in this industry since “21, I’m excited to see regulators, enterprises and startups all drive a new fin system together. There’s still much work to do for a new financial system, so we aim to flesh this out more and more, especially tied to relevancy of tech providers x jurisdictions. And beyond the financial system: I think we’ll move to @balajis “rule of code” eventually. Programmable compliance is definitely a part we need to bring that type of state to life. Let’s see where the world evolves into. Inshallah.

Is blockchain boring now? Kind of, yeah. Because we figured most of it out. The big inventions are behind us: from 2016 to 2021 we basically figured out every major primitive (high-throughput L1s, L2s, DeFi, AMMs, PMs, oracles). This is common in the history of technology. The most important ideas behind the Internet were figured out in the early 90s; the next 30 years were mostly just scaling it up. That's where crypto is now. What's left is implementation, UX tuning, and optimization. Which feels different, because it is. It's a sign of a mature technology. And it's a necessary phase in any technology that actually wins. I break down where we are in the crypto innovation cycle in this Stanford talk with @avichal 👇

Weird but true: without crypto, the current AI boom would not have been possible. Today's data center buildout would've been impossible without Bitcoin miners spending a decade building the knowhow around power, sites, and scale in remote locations. Nvidia could not moved beyond gaming to specialized compute without crypto mining as the bridge. And the biggest neoclouds are almost all ex-crypto talent: @CoreWeave, @CrusoeAI, @togethercompute, @nscale, etc.

My girlfriend has thousands of TikTok’s saved- day trips, vacation ideas, restaurants, gifts, etc. I couldn’t keep track of it. Now she texts the videos to Hermes, it watches them + stores in a vault, I text the agent to pull from her lists and plan. Magic.

Today we're shipping screen-aware dictation. First, we built a speedy speech-to-text (very fast, ~450ms). But, many products do this! So we went further. Now dictate using your screen as context. In Claude Code, it writes the prompt. In G-Mail, it replies in your voice. Demo:

one of my favourite GTM case studies has been was when Pravesh Mistry (head of global sales) joined Notion, the company already had product-market fit in enterprise. but enterprise wasn't proving to be a useful ICP; it was still too broad. every salesperson could interpret it differently, build a different pipeline and end up chasing completely different customers. so instead of expanding the market, they narrowed the entry point. they identified Engineering, Product and Design (EDP) teams as the wedge into enterprise. these were the teams actively looking for better ways to collaborate, had the highest product affinity, and most importantly, influenced how the rest of the company adopted software. the plan wasn't to sell Notion to the entire enterprise. it was to win the teams that would organically pull everyone else in. over the next ~3 years, enterprise ARR grew from ~$1M to ~$100M. win rates improved, deal sizes increased and sales cycles got shorter. what looked like a simple ICP exercise was really a distribution strategy. the best ICPs aren't just the customers most likely to buy your product; they're the customers most likely to bring everyone else with them.

most neobanks will not survive the next 18 months. not because demand disappears. $245M in top-ups in a single week proves demand is the least of your problems, they will die because of what they built underneath: i review compliance infrastructure for a living since 2017. here is the full map of what actually holds this market together, layer by layer, and who is powering each one right now cards your card program is a three-party compliance relationship: you, your issuer, and the network. the network's enhanced due diligence sits on top of your issuer's requirements. if either loses confidence in your stack, the card stops. not slowly. overnight @binance lost Visa in Europe July 2023. lost @Mastercard in latam two months later. gone by December. @ready_co gave non-EEA users one hour's notice in June 2026 when their issuer relationship broke. one hour what the network actually wants to see: account-level OFAC and sanctions screening, not batch, not periodic, continuous. a transaction monitoring system that produces real alerts. a KYC layer defensible across every jurisdiction you operate in. one audit trail running through every product the customer touches Starlingbank had a system that produced zero individual sanctions alerts for six months. £29M fine. that is the floor the infrastructure powering this layer right now: @raincards (Visa and Mastercard principal member, BIN sponsor for 200+ programs, one API for issuance, compliance, FX and onchain settlement), @pomelo_latam ($160M raised, powers bbva , santander , @Bancolombia , @WesternUnion , Binance across latam, just launched global stablecoin card across 150+ countries), @marqeta ($383B processing volume in 2025), @lithic, @GalileoFintech, @unit_co_, @treasuryprime, @Adyen, @Stablecoin @eldoradoio @Uglycash the compliance layer that makes the issuer relationship survivable is what @blend_money is built around: screening, audit trails, per-jurisdiction reporting, the infrastructure that keeps the card program intact at scale on and off ramps every ramp is a compliance event before it's a UX event on-ramp: you are opening a new account. source of funds, identity verification, risk scoring before a single dollar moves off-ramp: withdrawal with a clean audit trail, documented source of funds, per-jurisdiction reporting. this is where most teams underinvest because users don't see it. regulators do best practice: per-account screening on every transaction, not customer-level screening on signup and never again. your banking partner will pull a sample during their quarterly review. if the trail isn't clean per transaction, you find out at the worst moment the infrastructure powering ramps right now: @moonpay (eliminated fees on stablecoin onramps, enterprise stablecoin services live), @Transak (published the Q2 2026 compliance cliff report, most serious public documentation of what payment companies need before july), @Stablecoin (acquired by Stripe for $1.1B, trust charter approved february 2026), @belo_app @AlchemyPay, zerohashx, @Bitso , @RipioApp @daimo @dakota_xyz @RampNetwork @tazapay earn (im biased here just a little bit) the most misunderstood compliance surface in the stack shared vaults feel like a product architecture decision. they are actually a legal structure decision. commingled user funds create fiduciary exposure, insolvency complexity, and a direct failure point in any serious institutional diligence process the question that kills shared vault structures is simple: show me the ledger entry for user X's balance. if the answer requires reconstructing it from pool accounting, you don't have an answer best practice: isolated per user from day one. each account its own ledger entry. yield calculated individually. never commingled. this isn't conservative. it's the only structure that survives the question above from a banking partner, a regulator, or an institutional LP doing diligence on your cap table this is the architecture @blend_money runs. isolated accounts, clean ledger, never commingled for the institutional layer on top: @noon_capital brings the DeFi stack diversification and insurance coverage that makes yield products viable for institutions. diversified protocol exposure across @MorphoLabs, @eulerfinance, @pendle_fi, tokenized treasuries, CLOs and private credit. insurance gating on every deployment, no capital deployed without coverage. that is the version that survives institutional diligence the broader earn infrastructure: @opentrade_io (RWA-backed yield-as-a-service, bank-grade legal structure with bankruptcy-remote SPC, powers Littio, Kredete, Criptan), @OndoFinance, @maplefinance, @goldfinch_fi, @SuperstateInc in europe, MiCA Article 50 prohibits interest on euro-denominated stablecoins. the compliant path runs through tokenized T-bills and RWA wrappers. yield from an underlying asset, not from the stablecoin itself. whoever builds this first owns european earn cashback and rewards every rewards program with monetary value has reporting obligations (ps @itstuyo set the new standart here: buy now pay maybe) the cleanest structure: rewards funded from interchange revenue, paid in a regulated stablecoin, accounting that reconciles per user per period, tax-reportable from day one in every jurisdiction paying rewards in your own token introduces volatility risk for the user and securities classification risk for you. the question "is this a security?" becomes harder to answer the moment the token fluctuates and users expect returns compliance screening, the layer underneath all of it most teams assemble this reactively. something breaks, a regulator asks a question, a banking partner flags a transaction. then the compliance stack gets built. that is the wrong order the teams that survive build it preventively. before the card. before the ramp. before the earn product. one continuous audit trail across every product the customer touches the point tools doing parts of this well: @chainalysis (blockchain analytics, OFAC and sanctions screening, regulator-accepted in US, EU and UK), @elliptic, @trmlabs, @Sumsubcom (KYC, AML and Travel Rule in one integration, MiCA and FATF ready), @ComplyAdvantage, @notabene_id, @sardine, @unit21inc, @jumio, @Onfido but point tools create point gaps. your KYC vendor does not talk to your transaction monitoring. your transaction monitoring does not feed your sanctions screening. your sanctions screening does not generate the audit trail your banking partner needs to read. every gap is a reconciliation problem you find at the worst moment what @blend_money built is the integrated layer. AML screening, OFAC checks, KYC, transaction monitoring, per-jurisdiction reporting, all running together as preventive infrastructure before a single user touches a product. not a compliance dashboard bolted on top. the foundation the card, the ramp and the earn product sit on IDmerit's February 2026 breach of approximately 1 billion records made this clear: your compliance infrastructure is now a counterparty risk decision, not just a regulatory one the right order of operations 1) screening and transaction monitoring, then issuer relationship, then card 2) source of funds framework, then ramp, then volume 3) isolated ledger, then earn product, then institutional partners 4)interchange accounting, then rewards, then retention teams that invert this order ship faster in year one and rebuild in year two. sometimes year two doesn't come the $245M is not a card story. it's not a yield story. it's a survival story the neobanks still standing when this market hits $2.45B will be the ones that figured out compliance is not the last thing you build. it's the only thing that lets you build everything else WaveCrest taught this lesson in 2018. Wirecard taught it in 2020. Ftx in 2022. Binance in 2023. Ready in 2026 the lesson does not change. only the names do.