This edition is slightly later than usual because England decided to beat Mexico 3-2 last night in a World Cup knockout game that apparently required every nerve ending in the country to be personally tortured. So yes, blame Jude Bellingham, Harry Kane, and basic patriotic malfunction.

Back to the hiring market, which somehow managed to be just as dramatic. Microsoft is reportedly preparing another round of cuts while also launching a new $2.5B AI adoption company. Together AI raised $800M at an $8.3B valuation. Samsung is expected to post an 18-fold profit jump on AI memory demand. SK Hynix launched a $28B US listing. And India’s AI hiring is rising while broader IT recruitment falls.

The signal this week: AI is not killing hiring. It is making hiring more uneven. Specialist demand is rising, generalist demand is weakening, and infrastructure is eating the budget like a Victorian orphan at a buffet.

The Drop

1) Microsoft cuts while launching a $2.5B AI adoption machine

What happened: Microsoft is reportedly planning to cut under 2.5% of its workforce, with thousands of roles potentially impacted across sales, consulting, and Xbox. At the same time, Microsoft launched Microsoft Frontier Company, backed by $2.5B, to help enterprises adopt AI using Microsoft and non-Microsoft models.

Why it matters for hiring: This is the new enterprise AI pattern: cut where work is being commoditised, then invest in roles that help customers actually adopt AI. The “AI transformation” job is moving from slide decks into implementation, model selection, data integration, and measurable business outcomes.

Roles likely to spike:

• Enterprise AI implementation consultants

• Solutions engineers

• AI product managers

• Data integration engineers

• Change / adoption leads who can tie AI to measurable ROI

Roles likely to be squeezed:

• Generic consulting

• Lower-leverage sales support

• Non-technical project coordination

• Admin-heavy operational roles

2) Together AI raises $800M at an $8.3B valuation

What happened: Together AI raised $800M led by Aramco Ventures, more than doubling its valuation to $8.3B. The company helps businesses train and run workloads on open models like DeepSeek, MiniMax, and Kimi at lower cost than closed systems.

Why it matters for hiring: The open-model layer is becoming a serious enterprise choice. Companies want flexibility, lower inference cost, and less dependence on one frontier provider. That means hiring teams need people who can benchmark, deploy, secure, and monitor multiple model families.

Roles likely to spike:

• Open-source AI platform engineers

• Model evaluation engineers

• Inference optimisation engineers

• ML infrastructure engineers

• AI security and governance specialists

Hiring takeaway: “Experience with OpenAI” is not enough anymore. The better job specs will ask for model evaluation, routing, cost optimisation, and deployment across multiple providers.

3) Samsung expected to post an 18-fold profit jump as AI memory demand explodes

What happened: Samsung is expected to report Q2 operating profit of 86 trillion won, around $56.35B, up from 4.7 trillion won a year earlier. Analysts point to the AI memory shortage, rising DRAM and NAND prices, and demand from agentic AI systems that require more memory and storage.

Why it matters for hiring: AI is not just a model story. It is a memory, storage, supply-chain, and capacity story. If memory pricing keeps rising, AI teams will face more pressure to build efficiently.

Roles likely to spike:

• Hardware-aware ML engineers

• Performance engineers

• Inference optimisation specialists

• Capacity planning / FinOps

• Data platform engineers who understand storage cost

Hiring takeaway: Cost discipline is becoming a technical skill. The best AI engineers will not just ship features. They will know how to reduce compute and memory waste.

4) SK Hynix launches a $28B US listing to ride the AI chip boom

What happened: SK Hynix launched a US share sale to raise about $28.07B through Nasdaq-listed ADRs. The company has been one of the biggest beneficiaries of the AI memory cycle, with shares up roughly 260% this year. Proceeds are expected to support chip factory buildout and equipment purchases.

Why it matters for hiring: The AI talent market is becoming increasingly tied to semiconductors, memory, and physical infrastructure. Software recruiters who only understand “AI engineer” as one bucket are going to miss half the market.

Roles likely to spike:

• Semiconductor software engineers

• Hardware systems engineers

• Compiler / runtime engineers

• Data center integration specialists

• Supply-chain and chip manufacturing operations

Hiring takeaway: The AI stack is widening. The money is not just flowing into apps. It is flowing into chips, memory, power, cooling, and industrial buildout.

5) India’s AI hiring rises while broader IT hiring falls

What happened: Naukri data showed AI-related hiring in India’s IT sector rose 16% year-on-year in June, while overall IT recruitment declined 3%. Across 14 sectors, AI and machine learning jobs rose 25%.

Why it matters for hiring: This is the clearest workforce signal of the week. The market is not uniformly bad. It is bifurcating. General IT hiring is weaker, but AI-specialist hiring is moving in the opposite direction.

Roles likely to spike:

• AI / ML engineers

• Applied AI engineers

• Data science and model evaluation

• AI-enabled automation roles

• Senior specialists who can lead implementation

Hiring takeaway: Companies are not freezing hiring equally. They are becoming more selective, more senior-heavy, and more ruthless about whether a role links to AI capability.

AI Tool of the Week

SeekOut Recruit Core

What it does: SeekOut Recruit Core gives smaller hiring teams access to enterprise-grade AI sourcing, including a 1B+ profile talent index, Smart Match search from a job description, GitHub / academic / expert profile search, and AI-assisted sourcing workflows.

Who it’s for: Lean recruitment teams who need to source technical talent but do not have time to build 47 Boolean strings while slowly losing the will to exist.

Quick pilot idea this week:

• Pick one hard AI / infra role.

• Paste the job description into Smart Match.

• Build a first 50-profile shortlist.

• Manually review the top 20 for relevance.

• Compare it against your normal LinkedIn-only sourcing output.

Metrics to track:

• Relevant profiles per hour

• Outreach response rate

• HM approval rate

• % of candidates not found through LinkedIn Recruiter

• Time-to-first-qualified-shortlist

Hiring / Interview Insight

“AI-ready” hiring now means cost, evaluation, and workflow judgment

The mistake this week would be reading the news as “AI hiring is booming.” That is too simplistic, which is unfortunate because simplistic things are comforting and usually wrong.

The real pattern is:

• AI specialists are in demand.

• Generalist roles are under pressure.

• Infrastructure cost is rising.

• Open-model deployment is becoming more important.

• Enterprise adoption needs measurable ROI.

Add a 30-minute “AI operating judgment” station:

Give candidates this scenario:

“You’re asked to roll out an AI feature for enterprise users. You can choose between a closed model, an open model, or a hybrid routing setup. The feature must be secure, cost-effective, auditable, and reliable. What do you choose, what do you measure, and when do you escalate to humans?”

Score for:

• Model selection logic

• Cost awareness

• Evaluation discipline

• Security / privacy awareness

• Monitoring and rollback planning

• User workflow understanding

Metrics to track:

• Interview pass-through by seniority

• New hire 60-day manager satisfaction

• AI feature rollback rate

• Cost per successful AI workflow

• Time-to-productivity for AI-heavy roles

Funding Watch

Together AI | $800M round | $8.3B valuation

Open-model AI infrastructure for training and running AI workloads.
Likely hires: ML infra, open-model deployment, inference optimisation, enterprise solutions.

Crusoe | reportedly in talks to raise about $3B

AI data center startup reportedly in talks for a round that could triple its valuation.
Likely hires: data center ops, power, cooling, networking, SRE, capacity planning.

Oxmiq | $35M round | $60M total raised

AI chip architecture startup founded by Raja Koduri, aiming to reduce AI system cost by unifying graphics, CPU, and tensor engine architecture.
Likely hires: chip architecture, compiler engineers, IP licensing, hardware/software co-design.

SK Hynix | $28.07B US listing

One of the world’s largest new share sales, aimed at riding AI memory demand.
Likely hires: chip manufacturing, memory systems, process engineering, supply-chain operations.

Microsoft Frontier Company | $2.5B backing

New Microsoft-backed AI adoption firm helping enterprises select and integrate different AI models.
Likely hires: solutions, model evaluation, enterprise delivery, data integration, AI strategy.

Quick Bytes

• Meta’s Mark Zuckerberg said AI agent development is moving slower than expected, despite Meta’s huge AI infrastructure plans. This is a useful corrective to the “agents will replace everything by Thursday” crowd.

• Meta is projected to spend up to $145B on AI infrastructure this year, which is why every hiring plan now needs a cost-efficiency layer.

• Google’s dream-job status is reportedly weakening as AI startups offer bigger upside, more autonomy, and faster growth paths.

• Microsoft’s reported cuts are expected to hit thousands of roles, including sales, consulting, and Xbox.

• The Guardian reported questions around the UK’s Stargate AI infrastructure project, including whether parts of the investment story were more speculative than committed.

What to do this week

1) Split your hiring plan into “AI-leverage” and “AI-exposed” roles

Metric: every open role tagged as growth, neutral, or automation-exposed.
Why: Microsoft, India IT, and broader AI hiring data all point to uneven demand, not a simple boom or bust.

2) Add cost-awareness to AI job specs

Metric: every AI infra or product role includes one cost, latency, or evaluation expectation.
Why: Samsung, SK Hynix, and Crusoe show that AI infrastructure costs are not background noise. They are the market.

3) Build an open-model talent pool

Target profiles: DeepSeek, Kimi, Llama, vLLM, Ray, Kubernetes, model routing, inference optimisation.
Metric: 25 qualified profiles identified this week.
Why: Together AI’s $800M raise says enterprise AI is moving beyond one-model dependency.

This week’s takeaway is simple: AI is not producing a clean hiring boom. It is producing a talent sorting machine. The winners will be specialists who can make AI cheaper, safer, faster, and more useful. The losers will be teams still hiring for vague “AI experience” and hoping the job spec magically attracts someone who has actually deployed something.

That’s all for this week’s Tech Talent Drop — stay informed, and see you next week!