ELIS 2026: Still Alive. Still Kicking.

By Gabriel Karandyšovský and Monica Albini

In March, the European Language Industry Survey (ELIS) 2026 results were published.

For those unfamiliar with ELIS, it’s the largest pan-European research effort in the language services industry. Launched in 2013 by the European Union of Associations of Translation Companies (EUATC), it is today co-organized with ELIA, FIT Europe, GALA, the EMT university network, the European Commission’s LIND group, and Women in Localization.

ELIS is unique for its broad scope — it includes perspectives from universities and educational institutions, language services companies (LSCs, as ELIS refers to the more commonly used LSP term), buyer-side companies and organizations, and freelancers. It aims to capture the main trends and influences shaping it (note that ELIS collects data for the 2025 calendar year).

Marta Aragão, the president of the EUATC, shared in her opening remarks to the panel following the presentation of the results that ELIS serves a dual role: It is the industry memory and a means of broadening our collective intelligence.

We couldn’t agree more. Access to unbiased, inclusive information is essential for individuals and companies making informed decisions.

That’s why we at the AI Localization Think Tank decided to:

• Highlight and analyze ELIS’ main results and takeaways;

• Draw attention to the immense annual effort required to produce such a representative work.

So then, what’s the state of the language industry in 2026?

The first question that inevitably comes to mind: Is there anything else besides AI?

TL;DR

Here’s the short and sweet summary of the main ELIS findings to get started:

1. AI is the main driver of ongoing changes and industry transformation — but you already knew this. It makes for a tense reality, especially for solo-entrepreneurs (i.e., translators) or smaller LSCs, who are seeing their bread-and-butter services erode and their livelihoods increasingly at stake.

2. Demand for translation is on the decline, and AI is exacerbating price pressures (“AI is free, right?”), leading down a road where linguists simply cannot go.

3. The atmosphere in the room is loaded, and for good reason. That said, there are positives to be found (because, yes, that’s also a choice). For instance, now that we’re (mostly) past the AI hype and the world has seen the results of failed experiments and botched deployments, buyers realize that AI cannot do it all and are returning to humans.

4. Not all pockets of the industry are affected by AI the same way: For example, multimedia content is seeing a boom, while the demand for interpreting services remains as steady and as human-centric as ever.

5. Students have the right reflexes: They are already orienting toward pluri-disciplinary opportunities and acquiring skills across a broad range of areas to function and thrive in the AI world. It’s up to the universities to redesign curricula to keep pace.

6. Confidence in freelancing as a long-term career is declining across all experience levels, but the largest drop is among professionals with two to five years of experience — reality perhaps hits harder once you take off those rose-tinted glasses.

7. The current crop of students seems enthusiastic, but the talent pipeline doesn’t look promising: The industry needs to do a better job of attracting new talent.

It’s not fun to be an LSC in the AI-first world…

It’s not fun to exist in the “AI-first” world being foisted on us by tech companies and enterprises.

The sentiment is quasi-universal: AI (or the labs developing it, to be precise) is to blame. Sure, geopolitics adds unnecessary volatility, but on balance, ELIS 2026 confirms the general malaise of the language industry, especially from the perspective of freelance/solo-entrepreneur translators and smaller LSCs.

AI arrived at a point where MT/NMT had already been applying price pressure for years, and AI has only exacerbated it, forcing LSCs to adjust their workflows and review their service offering. This shows in the survey results:

• Prices have been trending downward for a few years now.

• LSC profitability dipped in 2024 as AI ate into demand, but rebounded slightly in 2025 as LSCs adjusted to it.

• Interestingly, LSCs report gross margins that are still in the healthy ranges (45% for translation, and 21% for interpreting) — a sign that the optimization (or cost-cutting, depending on your angle) LSCs needed to go through is working.

This contrast between dwindling activity levels and reported financials is noteworthy because it shows that the service providers still have levers to pull to offset dipping demand, which may not last and is, in any case, morphing.

…but better times ahead?

It’s interesting to contextualize the ELIS results against what market research firms such as Nimdzi or CSA Research report, despite a somewhat apples-to-oranges comparison and differing sampling.

For one, the top of the LSCs market (as reported in the Nimdzi 100, for example) seems to be in a more buoyant mood than the European SME respondents to the ELIS survey. Then again, in 2025, the UK Association of Translation Companies (ATC) reported that UK LSCs were more bullish about their reality and future prospects than their continental counterparts (this also shows in the sentiment reported by ELIS).

Who’s right? And what’s the underlying story?

There are multiple variables that significantly influence reality and future projections of LSCs:

• Their geography: Some markets have more structured and consistent demand for language services (e.g., through regulation or a strong public sector, as is the case for the UK)

• The scope of their service offering: Demand for “classic” translation is down, and AI-powered translation and AI-powered services (e.g., automated post-editing) are up; companies with a strong creative side to their service lines (e.g., multimedia localization or transcreation) are able to articulate their added value slightly better

• Their client mix: Clients in regulated industries are relatively safe from the adverse effects of AI, as well as those in more creative industries, where the human touch is still prized

Yes, the fortunes of the very top of the market are different than those of the SMEs making up the bulk of the landscape. But that doesn’t mean LSCs of any size are bereft of options for positively affecting their bottom line. More importantly, they are already taking the steps to help themselves (thankfully, the proactive attitude has sunk in).

The underlying story that the data only half-show is one of resilience. When faced with existential choices, LSCs have been largely showing up.

To ELIS’s credit, the presentation of the results rightfully sought to go beyond moroseness and toward pragmatism and concrete steps that individuals and companies can take to right the ship and/or carve out their own niche:

• Niche specialization has never been more viable, but diversification is valuable too — it depends on the company's size and the resources available to pivot.

• Direct client work is infinitely more rewarding (not just monetarily) than working through other LSC intermediaries, especially at a time when buyers are coming to prize the partnership LSCs offer, whether that shows through the subject-matter/technical expertise they provide or the additional bandwidth for clients who have none.

• At the end of the day, AI is a tool. There is a collective realization of just that. This means that LSCs who can use it as a tool to multiply their capabilities — or use AI internally to identify and/or remove waste — are having an easier time transitioning to the new reality.

So, what’s this new reality?

Stepping into the buyers’ shoes

The ELIS results touched on this, but we can also couch our commentary in observations from the conference circuit, or from working with buyers in our day jobs: When talking to buyers, it is evident AI is here to stay. The C-level wills it, and there’s very little that can be done about it.

The conversations clients are having are now moving into a different space: One about risk management/control, and how best to integrate technology and redesign traditional workflows.

ELIS reported on this, too, which should be a positive for the service provider portions of the crowd:

1. Now that buyers are generally moving from experimentation and implementation to scaling, the lack of resources and technical knowledge is felt — and LSCs are clearly positioned as the go-to partners for integration.

2. Secondly, as buyer-side organizations explore the limits of AI and the risks it poses, creating bespoke, secure solutions or embedding AI within secure work environments is essential. Here, too, LSCs have a role to play, helping clients to establish AI governance principles and mitigate its harmful effects.

These currents mirror what you’ve surely read elsewhere or experienced in your business: The move for linguists and LSCs is away from a relatively narrowly defined doer to a more consultative partner.

Information to action

Here, it is opportune to circle back to the importance of ELIS in the broader industry information landscape: Access to quality, unbiased market intelligence has never been more important.

Especially for the next generations.

Students continue to rely primarily on teachers and newsletters for industry information — only 25% turn to additional sources, such as guest lectures or direct industry contacts. The reported reason? Lack of time.

In the meantime, active participation in professional events and forums is declining (perhaps due to the cost barriers they pose for students and early-career professionals). University staff, for their part, flag a persistent problem: Access to market information and industry expectations is limited.

What makes this picture particularly striking is that students do show alignment with the market. Their expectations around market size are closer to those of LSCs than to those of their own teachers, and also their employment expectations align with employers’ staffing indications (unfortunately, in a pessimistic direction). The ELIS report also highlights the role of internships in building awareness and judgment, suggesting that hands-on exposure may be one of the most effective bridges between academic training and professional reality.

For master’s programs, internships are described as a core part of training: 72% are compulsory, and 23% are optional. But the ELIS report also shows growing strain on placement availability: Approximately 30% of LSCs are not accepting interns, especially due to a lack of work and resources to support the interns. Consequently, companies are becoming less willing to host interns, and some university departments are accepting fewer placements, thereby straining practical training opportunities. For buyer-side organizations facing AI-induced budget and staff cuts, the idea of interns seems even more far-fetched (though some companies, such as Adobe in the US, have historically on-boarded student talent).

There was a question posed during the presentation of the ELIS results that remains distinctly open-ended: Are there enough internship opportunities in our industry these days?

Student focus is evolving

Interestingly, despite AI's adverse effects on the industry, universities are not seeing a collapse in student numbers, especially in bachelor's modules. There is a concern about the talent pipeline, though: Evolving student interests and significant structural challenges in training and AI integration are problematic. The question of how to attract future talent is real.

At the same time, university staff worry that fewer students are choosing translation-related programs, a concern reflected in a visible shift toward combined and more applied curricula. ELIS reported that 35% of participating students chose a combined Translation & Interpreting program, up from 21% in 2025, suggesting growing interest in mixed or broader training paths. Students also reported increased popularity of disciplines such as creative translation, medical translation, project management, and technical writing. In other words, student interest in language careers remains stable, but there's a marked shift towards combined programs and growing demand for specializations.

This begs the question: Are students already choosing to pursue mixed/combined disciplines better prepared for the future than professionals already in the field?

AI integration in education: WIP

AI's role in language education is still developing. Only 47% of training staff report formal university-level rules on AI use, indicating that the consistent integration of AI into translation training remains a significant work in progress. University staff and independent professionals share broadly similar views on sector trends, both leaning toward skepticism about AI impact but anticipating a post-hype market correction.

It’s not all bleak, however. Positive sentiment is largely tied to ethical AI applications or, in some students' cases, to active resistance movements and the expectation of renewed market clarity once AI enthusiasm subsides — something which is slowly materializing, starting on the client side.

It’s interesting to note that both trainers and students identify signs of resilience and adaptation, though from slightly different viewpoints:

• Trainers tend to look at the profession itself: They see opportunity in the productive integration of technology into existing workflows, a growing and more nuanced understanding of AI's actual capabilities, and also a gradual evolution of professional roles to meet new demands.

• They also point to something crucial: Renewed recognition of human expertise, suggesting the industry may be finding stability after an initial period of disruption.

• Students, on the other hand, have a broader lens, highlighting the opportunities that globalization, technology, and social change are generating. They echo the trainers' sense of a grounded "AI realism", and point to expanding sectors and increased global mobility as drivers of a higher demand for translation and language services.

• Yet alongside this cautious optimism, students are clear about what they feel is still missing: More industry-relevant content, stronger training in what it means to be freelance and in entrepreneurship, regular AI updates, better communication, and a broader reformulation of existing curricula with an eye toward professional life after graduation.

Toward new versions of our (professional) selves

ELIS 2026 was the most recent dose of data on an industry that is clearly being reshaped by AI.

There are decisions to be made across the board and irrespective of the role in the landscape: proactive vs. reactive, close shop vs. find pathways to evolve and rediscover success, be more aware of your career choices instead of enrolling simply because you like languages (which you still can do, of course), and more.

AI has not stripped us of our agency. On the contrary, it has reminded us that we, humans, have choices to make.

For the service provider side, this may mean enduring a lot of discomfort in the short term, but staying resilient and building toward a different, more evolved version of their DNA.

For clients, the nature of their roles is morphing. The language work they do pass on to the industry is not diminishing. If anything else, it is acquiring new dimensions, such as risk (tied to quality) and security, that need to be managed.

Students still see language careers as meaningful, and their expectations are closer to market reality than to academic projections. But this alignment is fragile: Access to industry information remains essential, internship opportunities are few and far between, and the curricula students are asking for are not yet on offer.

Universities need to prepare the next generation — not just as linguists, but also by teaching business acumen, building awareness and realistic expectations, and equipping them with the tools to thrive in a profession where AI will be the baseline.

If this sounds like a lot of work for all of us, yes, it is. This is the job.