Graduate signing windows are closed. Q1 headcount additions are on the books. Compensation resets that were "under discussion" in January have either landed or been quietly shelved. What remains is the data — and in H1 2026, the data describes a market that bifurcated further, faster than most CHROs planned for.
CompTIA counted 275,000-plus active US job openings referencing AI skills in January 2026 — a 153 percent jump from January 2024. By March, the BLS JOLTS dataset showed 6.9 million total US openings against 4.8 million hires, a gap that has widened each quarter since Q3 2024. The AI-specific gap is sharper still. Machine learning engineer postings are up 59 percent since early 2020, per Indeed Hiring Lab, while general software engineering postings are down 49 percent. Two labor markets are running in parallel. One is tightening. One is not.
This is the launch issue of "State of Hiring · H1," the monthly theme running through June. Each bureau files below.
US: OpenAI's 100-Hires-Per-Week Pace, Anthropic's Efficiency Thesis, xAI's Reset
OpenAI announced in March 2026 that it intends to reach 8,000 employees by year-end — nearly doubling its headcount from approximately 4,500. Bloomberg and the Financial Times both confirmed the plan, which requires OpenAI to onboard roughly 3,500 people across nine months, or close to 100 net new hires per week. The growth is distributed across product, engineering, research, and a category the company is calling "technical ambassadorship" — specialists embedded with enterprise clients to accelerate deployment.
OpenAI is absorbing headcount at a pace that tracks Anthropic's competitive pressure as much as its own model release calendar. The March memo authorizing the expansion opened, per two people briefed on its contents, by noting that OpenAI's current organizational density does not match its revenue trajectory — which the FT has estimated at $11.6 billion annualized entering 2026.
To fund 1 million-plus square feet of San Francisco office space and a 3,500-person hiring campaign simultaneously requires a revenue base that most companies at OpenAI's stage would not have. OpenAI has it, and is spending it.
Anthropic is running the opposite thesis. The company's headcount sits at approximately 2,300 full-time employees — a figure that SaaStr flagged as extraordinary given Anthropic's revenue scale — with roughly 400 open roles posted at any given time. Anthropic's go-to-market share of those open roles has risen from 17 percent to 31 percent since Q1 2025, per Epoch AI job posting analysis, which explains where the growth is actually happening: not in the research lab, but in the commercial layer being built beneath it.
The compensation floor at both companies has reset upward. Senior software engineers at Anthropic (L5 and above) land $316K base plus $247K in annual equity, per Levels.fyi 2026 public submissions — total comp around $600K at the median senior IC level. OpenAI's L5 benchmark is higher: $336K base, $774K in annual equity, $1.15 million total. Research scientists at OpenAI cross $925K at the median. At the 90th percentile across both labs, total comp clears $1.28 million. These are not outlier packages for negotiated star hires. They are the published midpoints.
xAI's situation is distinct. The company registered 5,000-plus employees on LinkedIn entering 2026, but underwent restructuring following the X Corp integration — consolidating into four primary engineering organizations. A TechCrunch investigation published in March 2026 described internal architecture decisions being relitigated: "not built right the first time." Grok's internationalization hiring — roles for Hindi, Bengali, Russian, Arabic, Mandarin, and Indonesian speakers posted in January 2026 — suggests xAI's near-term hiring priority is model capability breadth rather than headcount expansion. Compensation for those roles ranges from $120K to $200K-plus in base salary, well below frontier-lab norms for AI research.
The week's specific movement: xAI opened additional roles in its Memphis compute infrastructure cluster. Anthropic extended offers to at least three senior policy staff from former Inflection talent networks, per two people with knowledge of the placements. OpenAI posted 43 net new roles in the seven days ending May 30, skewed toward enterprise sales and solutions engineering — consistent with its "technical ambassadorship" build-out.
Middle East: Saudi Arabia's $100 Billion Bet, UAE's Hiring Acceleration
Saudi Arabia's Public Investment Fund formally launched Humain in H1 2026 — a $100 billion AI fund structured as an independent operating company, not a passive investment vehicle. Humain is placing orders for hundreds of thousands of chips and targeting a top-three global position in compute supply. Humain requires a technical workforce with no domestic analog: it is recruiting internationally against US lab compensation benchmarks while offering something no US lab can replicate — the chance to build sovereign AI infrastructure at national scale, from scratch, on a decade-long mandate.
Saudi Arabia faces a 50 percent hiring gap in advanced AI roles, per the ManpowerGroup GCC Talent Shortage Survey (2026) and corroborating estimates from multiple GCC-based AI recruiters active in the Riyadh market. It is filling that gap with aggressive relocation packages to international AI engineers and a parallel investment in domestic training programs through SDAIA, the Saudi Data and AI Authority. The tension between those two tracks — import talent now, build domestic supply over time — will define Saudi AI hiring through 2028.
The UAE is moving faster on hiring velocity. UAE AI role postings grew 48 percent year-on-year in H1 2026, the highest growth rate among any country tracked by GCC workforce analysts including LinkedIn Economic Graph data and PwC Middle East hiring surveys. Demand for data scientists is up 43 percent year-on-year. AI engineer hiring grew 31 percent between 2024 and 2025, and the Q1 2026 data suggests acceleration rather than plateauing. Compensation for AI engineers in Abu Dhabi and Dubai sovereign-adjacent roles runs AED 380K–440K base, plus housing supplements of AED 60K–84K annually and employer-covered Golden Visa processing — a total first-year guaranteed value of AED 440K–524K, tax-free, with performance bonuses and milestone allowances reaching AED 570K in higher-responsibility roles.
The Abu Dhabi sovereign AI complex — ADNOC Digital, Core42, Presight, M42 — is now the single largest institutional AI hiring program in the Gulf, absorbing MBZUAI graduates, international senior hires, and mid-career engineers looking for the employer of record stability that GCC startup hiring has historically lacked.
Europe: August Deadline Pressure, Compliance Hiring Surge
August 2, 2026 is the enforcement activation date for the EU AI Act's high-risk and transparency chapters. National authorities across the EU gain full inspection and sanction authority on that date. Penalties reach €35 million or 7 percent of global revenue, whichever is higher. The compliance workforce required to meet that deadline was not built during the AI Act's drafting phase. It is being built now, at emergency velocity, by companies that underestimated the lead time.
The hiring consequence has been visible since Q1. Berlin holds the highest density of open junior AI governance roles in Europe — SAP's Walldorf organization, Siemens Energy's AI compliance function, and the German Mittelstand's enterprise AI stack are all posting simultaneously. The profile in demand is a graduate-level hire with both technical AI literacy and demonstrated EU regulatory fluency: a combination that did not exist as a job category in 2023. Salaries for AI compliance engineers in Germany run €75K–€95K base at the junior level, with senior staff roles in €120K–€145K territory.
Mistral entered this environment as a hiring competitor, not just a compliance subject. Its Programme d'Excellence en IA — launched Q1 2026 — enrolled approximately 22 graduates at €95K base with a conversion track to full-time employment. Hugging Face's EU taxonomy function is expanding to address August obligations for its open-source model releases. Both companies are simultaneously building products, managing compliance obligations, and competing for the same technical-legal talent pool as the enterprise incumbents.
The 45-plus major European companies that sent a joint letter to the European Commission requesting an implementation delay did not receive one. The August deadline is firm. Companies that planned for it are recruiting from a depleted talent pool; the ones that did not are in a worse position.
UK: London's Second Capital Moment
OpenAI's London office is now the company's largest research hub outside the United States. Anthropic's 158,000-square-foot King's Cross headquarters — planned capacity of 800 people, approximately 600 net new London roles across the build-out — lands in the same block as Google DeepMind, Meta, Wayve, Synthesia, and Isomorphic Labs. Any AI employer entering London today inherits a recruiting funnel of more than 10,000 AI-trained engineers within commuting distance. That density did not exist three years ago.
Google DeepMind employs approximately 2,000 staff in the UK — its global LinkedIn headcount sits at roughly 4,400 — and has anchored the King's Cross cluster since establishing its headquarters there. But the cluster's center of gravity is shifting as US frontier labs expand. Anthropic is offering some London engineers up to £630,000 annually — a figure first reported by City A.M. — which sets a new published ceiling for the UK AI market. OpenAI's London compensation is structurally similar, if less publicly stated.
The UK government's AI Opportunities Action Plan, refreshed in early 2026, designates Britain's goal as the third global AI hub after the US and China. The policy instruments supporting that goal — compute infrastructure investment, dedicated AI growth zones, UKRI industrial partnership funding — are real but lag the pace of private capital flowing into the London cluster. The talent race in King's Cross is moving faster than the policy framework around it.
Metaintro data cited in May 2026 counted 1,300-plus active AI roles in London, a figure that has risen each month since Q3 2025. The concentration is not evenly distributed: it sits in the Knowledge Quarter between King's Cross and Euston, and it is pulling senior researchers from Cambridge, Edinburgh, and — increasingly — US remote workers who want London-market exposure without UK-market compensation haircuts. The US labs are offering both.
The H1 Number That Sets Q3
The single figure that matters most entering Q3: 275,000-plus active US openings referencing AI skills, against a talent pool in which CS graduate unemployment runs at 6.1 percent and general college graduate unemployment runs at 4.1 percent (National Association of Colleges and Employers, NACE Class of 2026 Outlook Survey). The market is not short on candidates. It is short on candidates who clear the specific bar that frontier AI employers have set — and that bar has risen each quarter for six consecutive quarters.
The employers who moved fastest in H1 — OpenAI on headcount, Anthropic on go-to-market build, Humain on sovereign scale, Mistral on compliance positioning — did not move fast because the market was easy. They moved fast because the alternative was watching a narrowing pool of qualified candidates commit to someone else's offer.
The second half of 2026 will test whether the headcount additions made in H1 convert to the research output, revenue growth, and product velocity that justified the spend. That is the question this bureau will track through Q3.
The week opens here.