Phoenix Court

AI predictions: what to look out for in 2026

While the future is never certain, there are signposts of where the next tech hotspots will emerge to capture the attention of consumers, business and investors alike. Ziv Reichert, Phoenix Court's investment partner, shares his insights on the trends to watch in the coming year.

“Here are a few thoughts on what we might see in 2026. Some of these will likely happen sooner than expected. Others will not happen at all. The world moves too fast for predictions to hold.” - Ziv on X.

In the piece below he discusses the significant shifts occurring in software, the role of AI in transforming the future of work, the evolution of the content and media landscape as tech giants eclipse traditional Hollywood, and the influence of deeptech on healthcare and education.

Software creation becomes universal

  • Building software becomes the responsibility of every person in every function in every organisation.
  • We see a surge in custom software being built. Much of it runs locally, is rarely shared, and is designed for a single use case at a moment in time (see @tobi here).
  • Open source explodes as the number of people writing code force multiplies. GitHub has its best year ever for user growth. New GitHub competitors emerge to serve vibe coders.
  • Many non-open-source software companies make their products extensible so users can build on top of them (think @attio Developer Platform).
  • We see a resurgence of non-templatised, truly personal websites.
  • The bulk of design takes place in code. Newfound appreciation for well-designed design systems emerges as the number of contributors to codebases grows and lack of guardrails leads to frankenstein UIs.
  • We see more apps built on top of local models. We also see Claude Code type agents running locally on Opus-4.5 quality models.
  • Vibe coding comes to mobile. It starts to eat into TikTok and Reels engagement.
  • Vibe coders start tinkering with hardware.
  • We see an explosion of vibe coded games.
  • A significant share of new software discovery happens through agents.

Coding agents reshape work

  • ROI from coding agents becomes indisputable. Large companies face a difficult choice: layoffs or deadweight. Many choose layoffs.
  • Coding agents subsume the bulk of workflow automation software.
  • Long-running agents reliably complete multi-week software tasks (Metr, Cursor).
  • Learnings from coding agents (filesystem access, skills, rules, subagents, etc) get applied to many non-coding domains (Claude Cowork).
  • Markdown files become the standard for giving context to agents. What engineers are doing now with skills, rules, hooks, etc spreads across the enterprise. Organisations encode their institutional knowledge and processes the same way. Companies like Obsidian ride this wave.
  • Systems of record companies reorient around agentic users.
  • IDEs look wildly different by year end. Code shifts backstage. The focus moves to managing agents. IDEs reorient around helping users understand what their agent fleets are working on and which agents need their attention. The CLI gradually gets subsumed by these more human-oriented interfaces. Major companies (e.g. OpenAI, Anthropic, Cursor) begin to coalesce on a unified view of what the 'new IDE' looks like.

Agile dies

  • Agile and the product sprint as we know it die. New ways of building products emerge in response to coding agents.
  • The open office layouts we've grown used to get reimagined as voice becomes the primary way people interact with their computers (e.g @WisprFlow).
  • Long-term product planning collapses. Teams ship at the speed of inference.

Transcription and voice continue to diffuse

  • Transcription continues to proliferate across the enterprise. The bulk of internal and external meetings are recorded. Transcription remains a grey area with no widely accepted norms around consent.
  • People start to figure out novel ways to derive hidden value from their transcripts (think Granola Crunched). Workers increasingly push to maintain ownership over their transcripts when they change jobs.
  • Transcription seeps into people's personal lives as they grow accustomed to it in the workplace. Doctor visits are the first mainstream consumer use case.
  • Voice becomes the core interface for interacting with AI. This rewires how we interact with all of our products. Typing becomes a chore. Voice notes finally gain widespread acceptance on WhatsApp and other messaging platforms.

Roles and responsibilities blur

  • Competition in software intensifies as barriers to building keep falling. As a result, distribution becomes the responsibility of all employees, from entry-level to C-suite. The archetype for a star employee in 2026 is someone who A) has a real grip on what people are talking about online B) can actually ship product in response C) can create compelling content about what they've shipped, and in doing so builds an authentic brand synonymous with their company (@bcherny at Anthropic who led the build of Claude Code).
  • We see continued demand for FDEs, product designers, PMs (who can vibe code). And of course researchers.
  • Product designers start selling paywalled access to "artisanal" components that people can feed their agents.
  • Curiosity, agency, product intuition and ability to build become the highest value skills.
  • Brian Chesky's "Founder Mode" evolves to Sebastian Siemiatkowski's "Founder Code". All founders begin or go back to shipping code, irrespective of company size. The distance between founder vision and what's in production shrinks massively.
  • Companies stop referencing Github profiles when assessing engineering candidates as the platform gets inundated with AI-generated code. All engineering interviews require candidates to use coding agents. Many non-engineering roles start having coding interviews: product, design, operations, finance.
  • More VC dollars get spent on tokens than on talent.
  • The number of acqui-hires surges.

The media landscape shifts

  • X Articles replace Substack as the go-to way to publish online.
  • Traditional news outlets continue to lose their audiences to social media, namely to X. OSINT accounts rapidly replace newsrooms and citizen journalism explodes. We see many more breaking stories come to light that lead to real-world consequences (thinkNick Shirley exposing fraud in Minnesota).
  • X continues to cement itself as the preeminent place to understand what's happening in the world. "Is this real @Grok?" is the new fact-check.
  • Trading volumes on prediction markets continue to grow. The platforms increasingly intertwine with the news ecosystem. Traditional media begins to regularly cite market probabilities. xAI attempts to buy Polymarket or Kalshi. We see a cottage industry of forecasters emerge + a new cohort of hedge funds looking to trade these markets using proprietary and non-proprietary tooling.
  • AI-generated photos and videos (à la Grok Imagine) become the dominant meme format and how the internet commentates on world events. Everything's remixed with AI (think Maduro DJ). The value of the iconic shot compounds as a result.
  • The long tail of society begins to understand that most of the content they're consuming is AI generated. Nothing is perceived as real unless verified. The big platforms lean into AI content rather than away from it. Other platforms emulate X's 'Ask Grok'.
  • More developers begin distributing via the feed (e.g Wabi, the first personal software platform).
  • Influencers get into the software business.
  • Software becomes a new meme category.
  • New human-only social platforms emerge but fail to gain traction. Instead, people vibe code bespoke social apps for their immediate micro networks.
  • Vibe coding becomes a big streaming category on Twitch, TikTok Live and YouTube.

Hollywood waves the white flag

  • Costs for video models go down while quality continues to improve. Hollywood leans heavily into AI. Major deals get signed with AI studios. Many creative agencies shut down.
  • Lower-latency inference for generative media and easy-to-use infrastructure providers like fal unlock a new wave of consumer apps.

More people pay attention to their health

  • Rapid advances in AI meets science push more people to pay attention to their health as optimism around near-term breakthroughs emerges (Bryan Johnson x Don't Die!). People take less physical risk.
  • Sales for wearables accelerate. They all begin baking in AI features with hopes of becoming the everything app for health.
  • Apple revamps Apple Health.
  • Clinics see a record number of proactive blood tests as people look to get more data to feed their AI companions for self-diagnosis (ChatGPT Health).

Backlash picks up

  • We witness lots of carnage due to insecure, vibe coded software. Exposed API keys, leaked PII, no auth boundaries. The number of people writing code far exceeds the number who can understand it.
  • Anti-AI movements grow rapidly and come centre stage in the political conversation as real impact on jobs is felt. The one thing camp 'pro AI' and camp 'no AI' agree on is that this stuff is the real deal.

The broader economy adapts

  • Higher ed goes into complete meltdown as they fail to keep up with the rate of change. New privately funded AI-native schools emerge alongside a new class of education products leveraging coding agents to deliver interactive, guided, hyper-personalised courses.
  • The gig economy shifts from food delivery and taxis to data collection.
  • Consumer-focused services businesses (accountancies, law firms) begin to feel the pain of AI as demand plummets and clients churn to serve themselves.
  • Consumers spending $300+ a month on AI products becomes the norm.

Follow Ziv on X

To learn more about Phoenix Court, our funds and investments, contact Paul Bishop, Investor Relations and Finance Paul@localglobe.vc